investigating-metric-anomalies
Pass
Audited by Gen Agent Trust Hub on Jun 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a standard observability and incident triage workflow. It leverages vendor-specific tools (e.g., query-metrics, query-logs) provided by PostHog for data analysis.
- [PROMPT_INJECTION]: The skill involves processing data from external sources like logs and traces, which constitutes an indirect prompt injection surface. However, the risk is assessed as safe because the agent's capabilities are limited to read-only diagnostic queries and the instructions emphasize data correlation rather than command execution based on log content.
- Ingestion points: Log data and trace spans retrieved via tool calls in SKILL.md.
- Boundary markers: None specified.
- Capability inventory: metric-names-list, characterize-metric-anomaly, query-metrics, query-logs, query-apm-spans (all read-only diagnostics).
- Sanitization: None specified.
Audit Metadata